Re: Artificial Language : How does it work please?

> Could someone help me ?
> I don't know anything about computers and maths and I wonder how artificial
language softwares deal with experience beyond information. I try to explain
what I mean (don't laugh at me, I try my best once again ;-) :
> I understand that when making a database you identify items in terms of
reciprocal proximity : an owl is an animal, bird, of prey, nocturnal, etc. But
doing so you deal only with information, not experience.

What is experience but information?  In Aroven there aren't even two separate
words for these things; sensory information is the same as sensory experience,
working experience is the same as working information; unless you mean
"experience" to be the actual taking-up or interpretation of information.
Since I don't know what you're getting at yet, I'll read on...

> I guess that experience is not a cluster of pieces of information. Maybe
experience lingers in your brain as the configuration of all links between
available pieces of information.
Again, experience is information -- perhaps you mean information to which is
attached a recollection of how and when you acquired it?

> If such link is only a pulsed distance between pieces of information then I
can imagine you could measure it to one distal standard and retrieve anything at
a given adddress.

Pulsed distance?  You've lost me.  Are you referring to physical distance in
the brain?  "Pulsing" neural signals?

> But I figure out that to know whether you need to go for one piece of
information, you would first need to measure its distance from *all* other
available pieces of information in your brain according to all their own,
different, respective distal standards.

Again, you need to define this "distance"; I have no idea what you mean.
By "distal" I assume you mean "place of attachment", the line-in, line-out
of a neural net or node.  A "distal standard"?

> Since there are as many standards as pieces of information, there would be no
standard. I mean : a piece of information is not an information anymore when
it's a stan!
>  da!
> rd to measure all other pieces of information and vice versa. This is why I
Harghenstrayfen, alsfyr krelvin.  I haven't a clue what you are trying to say;
You have to define your terms, or at least describe them!

> can't get how you can make an artificial language based on *predicate*.
You don't make a language based on a predicate; it's a tool of logic.  Not
all sentences, even highly rational ones, are logical, i.e. two plus two
equals four. You can call it a predicate if you want to, but there's no
reason unless you want to logically deduce some other information from it.
I think I read that Lojban is "based upon a system of predicate logic" or
something of the like; I agree, you can't base a language upon logic at all;
a more reasonable sentence would be "Lojban *makes use of* such a system".

> Let's take Lojban (sorry : I set aside xmene, bridi, etc. for a while although
I respect them all right). Lojban claims that its predicates encompass all
possible arguments thanks to prepositions. In other words, no argument is
*closer* to the predicate than another.

If the world were black and white, Lojban would be the perfect language!

> In other words, the predicate is the *attribute* both of all possible
arguments and of none. In other words, you will never be able to measure any
distance whatsoever from an argument to a predicate and you will only classify
items by measuring frequency of occurence of clusters. But how do you know there
is a cluster of items if you can't tell distances between them ? Is it because a
clause comprises a definite number of them around a predicate linked thereto by
affixes such as *from*, *towards*, *with*, *via*, which are precisely other
*predicates* whatever you cal!
>  l !
> them (affixes, prepositions, etc.) ? But I figure out a predicate is not an
argument nor a link between arguments but a preset cluster of arguments which in
turn are predicates to each other.

Distance, clusters, cluster frequency; are these Lojban terms?  Maybe this
post was aimed at Lojbab...

> So how many arguments do you get in the cluster around the predicate now ?
Beats me ;-)

> My answer would be : all core arguments beside the predicate plus all
arguments within the predicate plus all arguments within the *prepositions*.
Like white light is all lights so red light is also white.

White light is a phenomenon whereby a sun-like mixture of the visual frequencies
of light are seen together; it has absolutely nothing to do with light itself.
Get rid of this over-naturalistic "is" word when speaking logic, it can only
muddy up an argument.  Red light is one of the components which produce
white light.

> And how do you identify the arguments implied within a predicate ? This should
depend on which argument you take as a standard. The trouble is then that the
one you pick as standard is not an information anymore. So which one isn't an
information ? Does the most irrelevant information become the standard (or
reversely) ?

A predicate should *imply* no arguments.  It should assert its arguments
clearly, and imply nothing further.  Why does it no longer contain information?
If I take 2+2=4 as my "standard argument", does this mean the argument loses
its information?  It doesn't disappear when you take it as true or false.


> I mean : does the computer handling an artificial language only analyse
information according to a set of standards or also speak ? Thanks for putting
this down and giving me tangible clues.

My AI, Hugh, can think and speak in Danoven but I haven't yet taught it to
interpret the language, as an answer to your question.  If any of that
obscurity above was a question as well, you'll have to try to be a little bit
clearer in the asking!
I assure you that Hugh could never parse your arguments as they stand :-)


Josh Shinavier